“Opinions are like %#^&@*s: Everyone has one, and they all stink!”
I tend to lean toward a cleaner version, especially in polite company:
“In God We Trust. All Others Bring Data.”
Or, as Tom Cruise and Cuba Gooding said in Jerry Maguire:
Read the opinions of commenters on almost any mainstream tech article on DevOps and you can see why DevOps is hitting ‘the peak of inflated expectations’ in the Gartner Hype Cycle. There is intense interest in DevOps, and lots of debate, but so many opinions, and scarcely any proof.
No surprise then that, in the DevOps world, DevOps Metrics are hot. I mean really hot. I’m talking ‘ghost peppers on an infrared grill in an Arizona summer on the surface of the sun’ hot. Metrics are becoming increasingly important as an antidote to the opinions rampant in the DevOps world, both in public and in private. Because in the end, objective data is the killer antidote to the poisonous opinions of DevOps detractors, doubters, and naysayers.
We have an excellent series of PuppetLabs/DORA ‘State of DevOps’ research reports that started in 2014 and is now in its third edition. This year for the first time, Splunk was a proud sponsor of that research, alongside our great partner, Puppet. Stay tuned for more as the results are released soon!
And of course, ‘measurement’ is one of the four core areas in the CAMS (‘Culture, Automation, Measurement, Sharing’) paradigm created by a couple of my mates, early DevOps leaders (and all-around great guys) John Willis and Damon Edwards.
So I was thrilled to see the incredible turnout at DevOps Days London for an Open Space session I suggested and kicked off on ‘Metrics that Matter’; and to host multiple roundtables on ‘Data-Driven DevOps’ at the Delivery of Things World in Berlin.
There was substantial interest at both events in measuring three key areas where Splunk can help to show the real impact of DevOps:
- The velocity of the development pipeline (e.g. commit rate, idea-to-cash time, build times);
- The quality improvement of new applications (e.g. test coverage, error rates, incident rates);
- The impact of code on business KPIs (customer sat, time on site, revenue).
It was also reassuring to see keen interest at both events – from practitioners and managers – in measuring the less tangible human impacts of DevOps, in areas like employee satisfaction, staff retention rate, sick days taken, ‘home office’ days, 3am callouts, referral rates for new hires, etc.
This is the real core of DevOps: increasing collaboration, communication, integration to delivery better software, faster; the human side of DevOps, combined with the business impacts.
There was so much demand that the DevOps Days London organizers kindly asked me to present a short 5 minute preso (aka an “Ignite”) about Metrics on Day 2. Foregoing all parties and meetups, I slaved into the night to prepare it, and by all feedback so far it seemed to be very well received. You can see it below, or watch a video of my full DevOps Days London presentation:
I am excited to learn from my customers every day how important Splunk is to their efforts to adopt DevOps principles. Customers like NBC Universal are Splunking their DevOps workflows, using machine data to provide visibility into their ‘metrics that matter’. Rather than just putting a finger in the air and guessing which way the wind is blowing, they, along with many other businesses, are measuring the velocity of their workflow, the quality of their applications, and the impact of new code on business KPIs.
I am truly excited to see, from California to London to Berlin and beyond, organizations around the world are starting to leave their opinions behind them (pun intended). They are instead focusing on ‘Metrics That Matter’, using objective data to drive their DevOps decisions.
Won’t you join them?